With its Digital Art History Lab DAHL, the Frick Art Reference Library seeks to encourage the use of digital tools in Art Historical inquiry. By providing datasets, workshops, and access to tools, the DAHL hopes to facilitate the collaboration between art historians and scholars across a variety of fields. The datasets presented here focus on Dutch Golden Age artists and collections.
In the 1980s, Yale University Professor John Michael Montias began compiling art inventories of 17th century residents of Amsterdam. Montias shared this invaluable resource with the Frick Art Reference Library to maximize its access to other scholars of the Dutch Golden Age. In total, The Montias Database records 1,280 inventories dated between 1597-1681 and the 51,071 individual works of art that they contained. Since Montias’ gift, art historians have had access to the data in a user-friendly, searchable format on the Frick’s website. However, the format currently provided is ill-suited for data visualization and analysis. In order promote new scholarly projects, the Montias Database is presented here in the form of CSV and JSON files.
The database is presented in two files: art objects and inventories. The 1,280 Inventories from the Gemeente Archief, listed by inventory number and date, include information on the owners’ lives, the appraisers, some original text in 17th century Dutch, and additional commentary by Montias. The art database includes provenance, artists, and auction information for the 51,071 works of art. It also includes prices for individual objects, making it possible to analyze values. The markdown file fieldnames.md should help guide users through the database’s fieldnames, abbreviations, and identifiers in order to improve ease of access.
The Montias Database has been paired with a dataset compiled by the Frick Art Reference Library. This dataset contains 203 works of art that were once or are currently attributed to the 17th century Dutch artist Johannes Vermeer (1632-1675). Currently, only thirty-five paintings are considered authentic. With this dataset, art historians may uncover hidden connections of provenance, attribution, style, and more between the 35 authentic Vermeers and 168 other paintings, which range from forgeries and misattributions, to lost or contested works. The dataset was created with a spirit of exploration and experimentation, for those wanting a dataset with which to explore digital art history tools.
When possible, each entry in the Vermeer dataset includes information regarding alternate titles, rejected attributions, provenance, current location, date, image links, material, dimensions, signatures and inscriptions, citations, and exhibition history. This information has also been broken up into smaller datasets of works of art formerly attributed to Vermeer, lost works by Vermeer, an inventory of the Dissius Auction of 1696, and more. Some records are incomplete: Library staff will continue to research and enhance these records in the upcoming months.
Our datasets are available in CSV format and JSON. We have used the standard UTF-8 encoding for our CSV files. The data in each file is identical. Both formats are provided for your convenience.
These datasets are placed in the public domain using a CC0 License.
While this dataset is open under the CC0 license, we have included a few guidelines that we hope you will follow when using this data.
We have included links to the images in the Vermeer dataset whenever possible. Please note that these images may be licensed differently. Additionally, while the Frick Art Reference Library’s Photoarchive includes some of these images, they may be available for on-site consultation only.
While we have worked hard to provide the most complete dataset possible, information included in this dataset may not be complete or accurate. We are offering the datasets as-is and wish to make no warranties or representations of any kind. If you encounter any errors or issues, please contact dahlprograms@frick.org. We plan to update the datasets with new and revised information on a semi-regular basis. You are advised to regularly update your copy of the datasets to ensure you are using the best available information.
We cannot accept pull requests for the data at this time. If you find a mistake or have a correction or a suggestion for improvement, please contact dahlprograms@frick.org.
When possible, the Frick respectfully requests acknowledgement as a source. This will help to preserve the datasets. By citing our data or providing acknowledgement, you will help others to replicate, verify, and further explore your interpretation of our data.
Additionally, please share your work with us! We love seeing how people are using this data.
Use of this dataset does not grant or imply The Frick Collection’s approval, commission, or support of your work. The Frick Collection retains the rights to all of its trademarks, and they are not part of the dataset. If you transform or modify to the dataset, you must clearly distinguish the resulting work as having been modified from The Frick Collection’s dataset. If you create a derivative dataset from the Frick dataset, we ask that you consider releasing the derivative under a CC0 license, which mirrors the licensing of the Frick dataset.
Thanks are due to the CMOA, MoMA, and Tate for providing excellent examples for these guidelines.